3 points levelling system December 21, 2017 09:02PM |
Registered: 9 years ago Posts: 155 |
Re: 3 points levelling system December 21, 2017 09:22PM |
Registered: 12 years ago Posts: 5,794 |
Re: 3 points levelling system December 21, 2017 09:33PM |
Registered: 9 years ago Posts: 155 |
Re: 3 points levelling system December 21, 2017 10:23PM |
Registered: 12 years ago Posts: 5,794 |
Re: 3 points levelling system December 22, 2017 06:29AM |
Registered: 9 years ago Posts: 155 |
Re: 3 points levelling system December 22, 2017 07:45AM |
Registered: 12 years ago Posts: 5,794 |
Re: 3 points levelling system December 22, 2017 08:29AM |
Registered: 9 years ago Posts: 3,525 |
Re: 3 points levelling system December 22, 2017 08:48AM |
Registered: 9 years ago Posts: 155 |
Re: 3 points levelling system December 22, 2017 09:34AM |
Registered: 12 years ago Posts: 5,794 |
Re: 3 points levelling system December 22, 2017 09:49AM |
Registered: 9 years ago Posts: 3,525 |
Re: 3 points levelling system December 22, 2017 02:09PM |
Registered: 9 years ago Posts: 155 |
Re: 3 points levelling system December 23, 2017 01:52AM |
Registered: 7 years ago Posts: 1,007 |
Re: 3 points levelling system December 23, 2017 04:50AM |
Registered: 9 years ago Posts: 155 |
Re: 3 points levelling system December 23, 2017 05:20AM |
Registered: 13 years ago Posts: 1,465 |
Re: 3 points levelling system December 23, 2017 06:16AM |
Registered: 7 years ago Posts: 1,007 |
Quote
leadinglights
I think that what may be going on is that the four legs of the "squashed frog" are flexible and the metal you have added is also flexible. The normal requirement of the spring leveling is that the mount of the adjusting screws is firm and that the bed only moves when the load exceeds the upward push on the spring while your added metal part just added unwanted flexibility. My guess is that Prusa printers found a system that worked - the legs of the carriage bent rather than bending the bed. There may be other reasons that the four point leveling remains popular - maybe it helps with anti vibration, dynamic systems can be complicated and anti-intuitive sometimes.
Just a passing thought, is there anything else that may be holding a part of the bed down? Bed heater wires come to mind.
Mike
Re: 3 points levelling system December 23, 2017 06:30AM |
Registered: 13 years ago Posts: 1,465 |
Re: 3 points levelling system December 23, 2017 07:36AM |
Registered: 7 years ago Posts: 1,007 |
Quote
leadinglights
Quote
MKSA
Four points is popular because people simply don't think past the fact that as it has 4 corners, therefore it needs 4 post/screws ........
I find it difficult to think that Joseph Prusa would have used the four point or the squashed frog without reason.
Mike
Re: 3 points levelling system December 23, 2017 07:53AM |
Registered: 9 years ago Posts: 3,525 |
Re: 3 points levelling system December 23, 2017 09:12AM |
Registered: 8 years ago Posts: 507 |
Re: 3 points levelling system December 23, 2017 10:54AM |
Registered: 12 years ago Posts: 5,794 |
Quote
Tinchus
Im on stuborn mode now jajajaj. I will figure this out...
Ok, I have removed the antivibration pads. Printer now is on its table wich is flat and levelled. I have checked again if Z rails are parallel, this time useing an asimeto digital caliper just to be sure my measurements are precise. And yes, the are PERFECT parallel: 34.61 mm in each end of them (I got 34.63 mm in 1 of them... but that is 0.02 mm of differences wich is inside the error margin, and when repeated measurements I got readings between 34.61 and 34.63 ... so...
I re checked bed flatness.. it is flat. I can see bumps in the surface, and putting a steel ruler I can see it is flat, and in those bumps U can see, I can measure useing a gage something between 0.05 and 0.1 mm. In fact, what youn see in the picture, in the right, there is a brand new mk3 bed, I bought it this morning just to be sure my bed was not bent by the 4 points method in the past. And I got the same results while levelling.
Quote
Tinchus
So now I tried 3 points levelling again. I start setting the reference screw as in Digital dentist instructions. I set there a gap of 0.1 mm between noozle tip and the bed (measured with a gage, I dont use the paper method). The I move to the pitch screw and set there a 0.1 mm gap too (when I do this, the gap at the reference screw decreases a little, to 0.05 mm). If I move now the x carriage between these 2 screws, I can perfectly see that the line between them is really and nicely levelled, showing that the bed is not warped, at least in this area?
Then I go to the roll screw and set 0.1 mm gap there too. I check now the pitch screw again and did a very very small adjustment, then another very small adjustment in the roll screw again. After this if I check levelness all around the bed is really beatiful (the most beatiful is that doing this take 3 minutes!!!!)> in all places I get a gap between noozle tip and bed of something between 0.05 and 0.1 mm BUT in the top left corner of the picture, right near the corner (no further than 3 or 4cms away) the gap is of 0.25 EXACTLY (measure with a gage). Im preious post I reported this gap to be in the lower left corner. That seems to change depending on the steps I take to do the levelling. If I follow the step described by digital dentist, That bigger gaps moves to the top left corner. If I follow another steps described in other sites, that gap appear in the lowest left corner...
Only Idea comeing to my mind: The roll screw, is supported on a piece of metal I builted in order to get this mod (sonce prusas uses only 4 point method). The Aluminium bed comes with a hole in the exact center in order to be able to use it with a 3 point method too, so that hole I know is exactly in center. Question is: if the hole in the piece of metal is not exactly in the center (I measured it and can be may be 0.5 mm out of center), could be creating some trouble with the 3 points method?
Re: 3 points levelling system December 23, 2017 10:54AM |
Registered: 13 years ago Posts: 1,465 |
Quote
MKSA
......................................... hold the bearings with tie wraps,...................................................
Re: 3 points levelling system December 23, 2017 11:02AM |
Registered: 12 years ago Posts: 5,794 |
Quote
leadinglights
Example of non-intuitive:
Holding bearings with ty raps for the forces involved is BETTER than using a standard pillow block in a mechanism such as a 3D printer. The forces lifting the bearing will always be less than the tension of the ty raps so the play will be limited to the play within the bearing whiie a pillow block will have a tolerance of something like 15mm +0.01mm - 0.00mm adding that much play. Additionally, the ty rap solution will be some 10s of grams lighter, will be easier to replace and will be cheaper. Machinated aluminium pillow blocks are not an improvement.
Mike
Re: 3 points levelling system December 23, 2017 11:06AM |
Registered: 13 years ago Posts: 1,465 |
Re: 3 points levelling system December 23, 2017 11:26AM |
Registered: 7 years ago Posts: 1,007 |
Re: 3 points levelling system December 23, 2017 11:38AM |
Registered: 12 years ago Posts: 5,794 |
Re: 3 points levelling system December 23, 2017 11:51AM |
Registered: 13 years ago Posts: 1,465 |
Re: 3 points levelling system December 23, 2017 10:44PM |
Registered: 9 years ago Posts: 155 |
Quote
the_digital_dentist
Quote
Tinchus
Im on stuborn mode now jajajaj. I will figure this out...
Ok, I have removed the antivibration pads. Printer now is on its table wich is flat and levelled. I have checked again if Z rails are parallel, this time useing an asimeto digital caliper just to be sure my measurements are precise. And yes, the are PERFECT parallel: 34.61 mm in each end of them (I got 34.63 mm in 1 of them... but that is 0.02 mm of differences wich is inside the error margin, and when repeated measurements I got readings between 34.61 and 34.63 ... so...
I re checked bed flatness.. it is flat. I can see bumps in the surface, and putting a steel ruler I can see it is flat, and in those bumps U can see, I can measure useing a gage something between 0.05 and 0.1 mm. In fact, what youn see in the picture, in the right, there is a brand new mk3 bed, I bought it this morning just to be sure my bed was not bent by the 4 points method in the past. And I got the same results while levelling.
4 point leveling won't permanently bend an aluminum bed. Metals act as springs until you stress them to the point of plastic deformation. As lousy as four screw systems are, they can't put a permanent fold in a piece of aluminum. If the plate is deformed, it started that way.
Quote
Tinchus
So now I tried 3 points levelling again. I start setting the reference screw as in Digital dentist instructions. I set there a gap of 0.1 mm between noozle tip and the bed (measured with a gage, I dont use the paper method). The I move to the pitch screw and set there a 0.1 mm gap too (when I do this, the gap at the reference screw decreases a little, to 0.05 mm). If I move now the x carriage between these 2 screws, I can perfectly see that the line between them is really and nicely levelled, showing that the bed is not warped, at least in this area?
Then I go to the roll screw and set 0.1 mm gap there too. I check now the pitch screw again and did a very very small adjustment, then another very small adjustment in the roll screw again. After this if I check levelness all around the bed is really beatiful (the most beatiful is that doing this take 3 minutes!!!!)> in all places I get a gap between noozle tip and bed of something between 0.05 and 0.1 mm BUT in the top left corner of the picture, right near the corner (no further than 3 or 4cms away) the gap is of 0.25 EXACTLY (measure with a gage). Im preious post I reported this gap to be in the lower left corner. That seems to change depending on the steps I take to do the levelling. If I follow the step described by digital dentist, That bigger gaps moves to the top left corner. If I follow another steps described in other sites, that gap appear in the lowest left corner...
Only Idea comeing to my mind: The roll screw, is supported on a piece of metal I builted in order to get this mod (sonce prusas uses only 4 point method). The Aluminium bed comes with a hole in the exact center in order to be able to use it with a 3 point method too, so that hole I know is exactly in center. Question is: if the hole in the piece of metal is not exactly in the center (I measured it and can be may be 0.5 mm out of center), could be creating some trouble with the 3 points method?
Quote
the_digital_dentist
Questions: Are you making leveling adjustments while the bed and nozzle are hot (you should be)?
I know I should. So far im just trying to see if the system will work. So im tsting in cold. Once I see th system works in my setup, I will level again, this time with bed ant hotend at printing temp
Quote
the_digital_dentist
Are there four bearings attached to the carriage plate, or three?
3 bearings, good quality, no play on them. 2 located in the right of the psted picture, 1 on the center left
Quote
the_digital_dentist
How about the extruder carriage? 4 bearings are as big a mistake as 4 leveling screws.
4 bearings in the extruder X carriage
Quote
the_digital_dentist
Does the plate rock on its bearings if you try to wiggle it? What about the extruder carriage?
Solid. No play on the bearings.
Quote
the_digital_dentist
When you have the printer moved to the corner that appears to be low, is the extruder carriage stable and is the bed stable?
yes, stable both of them
Quote
the_digital_dentist
What is the surface material on the bed? Is it uniformly thick?
that is kapton tape, only 1 layer in all bed
Quote
the_digital_dentist
How much do the X axis guide rails sag? You have a "low" corner, which may actually be a high extruder. The guide rails will tend to sag most when the extruder carriage is centered. That will make the bed look higher at the center than at the edges.
Yes, it sags a little (0.05/7 mm at the center, may be less). My extruder is very very light in order to try to avoid that and my smooth X rods are made a a hardned special steel I bought also to avoid this. My printer started as a cheap prusa i3 kit, but now, I have been replaceing every piece for hight qulity parts.
the "low" corner is the top left one. but if Y move and test levelness between center and top right corner, it is beatifully levelled. If X rails would be sagging, shouldnt I have the same oroblem with the top right corner?
Quote
the_digital_dentist
How about adjusting the roll screw with the extruder at a position about 1/2 way between the roll screw and the low, left rear corner of the bed. That should result in reduced "lowness" at the left rear corner and a little "highness" at the right rear corner.
will try this kind of adjustments tomorrow, yes
Quote
the_digital_dentist
If you adjust any one screw, the plate will pivot around the other two screws. Keep that in mind when you make the small tweaks at the reference or pitch adjusters. Also, the screw threads are typically pretty coarse. Small adjustments can have an unexpectedly large effect. Turning a 0.7mm pitch screw a half turn is going to move the bed up or down by 350 um. That's more than a print layer thickness.
the screws I use are stainless steel one, metric scres, very very small pitch, something like 0.2 pitch in order to have high precision
Quote
the_digital_dentist
The Taz that I converted to 3 point leveling at the makerspace is using the original, but modified carriage plate. The supports for the leveling screws are far away from the bearings, as in your printer. That makes leveling adjustments a little tedious because the act of inserting a tool to level the bed pushes the corner down slightly because the carriage plate flexes You have to insert the tool carefully, make a small adjustment, then remove the tool and measure to make sure that when the bed springs back up, the final position is where you want it. Repeat as needed.
I don't know how or why you're setting a 0.5 mm gap between the nozzle and the bed. The "paper method" works fine. There are folks who obsess over using feeler gauges because they don't know exactly how thick a piece of paper is, but it really doesn't matter. When I set the leveling and the zero point of the printer I use a piece of paper and set the Z=0 point to be where the nozzle just starts grabbing the paper. It works fine. Is a print's first layer exactly 0.200 mm thick or is it 0.220 mm thick? That depends on where you measure it and where the print was located on the bed. The sag in the guide rails (X and Y) and unflatness of the print surface (aluminum flatness + adhesive thickness + PEI thickness/flatness) guarantee that the first layer thickness is going to vary a little with XY coordinates. The Z=0 sensor always has limited precision, too. It's going to say that the bed or nozzle is at Z=0 at a different actual position every time you home the bed. That means measuring the exact gap between the nozzle and the bed is the wrong sort of thing to obsess over. Even if you set it accurately at a few points, it isn't going the be the same everywhere.
I'm as obsessive as anyone, but you have to temper that with a sense of proportion. There are things you can, things you can't, and things that don't need to be fixed. You can fix guide rail flex by using thicker guide rails or by using linear guides and supporting them along their length. You can't get nozzle to bed distance any more uniform than you can measure, and measuring things down to microns is extremely difficult, and pointless.
Don't lose sight of the goal: the only purpose of leveling the bed is to get that first layer to stick over as much of the bed surface as possible. With a reasonably flat bed such as cast tooling plate with a PEI layer, it's possible to print edge to edge, in some layer thicknesses. Nope, probably not edge to edge in 20 um layers, but yes, edge to edge in 200 um layers. The approximate thickness of the first layer can be tweaked easily by adjusting the Z=0 setting after you have the bed leveled and that first layer is sticking.
You're trying to get edge to edge performance from a piece of sheet metal. It may not be achievable. Like my rustic grandmother in Tennessee used to say, "you can't make a silk purse out of a sow's ear".
============ yes, I agree with all your comments. But If I have achieved a really nicel level in all the but but not in a small corner, that means Im doing something bad or something is in the setup. I dont expect to have perfect levelness as with a cast tooling plate, but haveing a levelness with an error of +- 0.2 mm is, I think and I have achieve better than that with 4 points method, what Im trying to get. With that margin, it is perfectly posible to compensate non levelness useing some overextrusion in the first layer.
I will report somethin tomorrow, thanks a lot for your comments
Edited 1 time(s). Last edit at 12/23/2017 10:51PM by Tinchus.
Re: 3 points levelling system December 24, 2017 10:22AM |
Registered: 7 years ago Posts: 1,007 |
Re: 3 points levelling system March 31, 2018 08:45PM |
Registered: 9 years ago Posts: 155 |
Re: 3 points levelling system March 31, 2018 09:38PM |
Registered: 12 years ago Posts: 335 |